Make it easier to select the used adaptation

Open nsiccha opened this issue 3 months ago • 1 comments

Currently, AFAICT, dual averaging is kind of hard-coded into NUTS, and the mass matrix adaptation simply depends on the type of the provided metric. Would be nice if there were a different way of specifying the used adaptation method, which "of course" should be orthogonal to the used metric.

Currently, this involved too much hacky boilerplate code for my tastes, see e.g. https://github.com/TuringLang/AdvancedHMC.jl/pull/473#issuecomment-3552348654

Nov 20 '25 09:11 nsiccha

The aim would be to also expose the changes in #473 when addressing this issue.

Nov 24 '25 12:11 nsiccha